Cognitive impairment is a prevalent, disabling, and costly condition in the elderly. Medication use is also common in the elderly, including those with cognitive impairment. The relationship between medications and cognitive function is important since medications may cause or exacerbate cognitive impairment especially when they are used inappropriately. The long term objectives of this proposal are to improve medication appropriateness among community-dwelling elderly individuals, focusing on drugs associated with the risk of cognitive impairment, and to advance pharmaceutical outcome research by conducting a comparative assessment of medication appropriateness. The risk of cognitive impairment in elders associated with the use of psychotropics, nonsteroidal antiinflammatory drugs (NSAIDs) or anticonvulsants will be assessed using three different designs: cross-sectional; prospective population-based cohort; and retrospective nested case-control. The specific hypothesis to be tested is that the use of psychotropics, NSAIDs, or anticonvulsants is associated with cognitive impairment. The reliability and validity of three alternative methods of assessing medication appropriateness for psychotropics, NSAIDs or anticonvulsants will be determined. The standard patient level-drug utilization review (DUR) and a population level DUR will be compared. The patient level DUR will also be compared to an alternative approach that utilizes standardized implicit criteria, the Medication Appropriateness Index (MAI). The specific hypotheses to be tested are that the patient-level DUR will be more accurate when compared to the population level DUR and that the MAI review will be comparable to the patient level DUR. The studies capitalize on data from the NIA-funded longitudinal Duke Established Population for Epidemiological Studies of the Elderly (EPESE) and a Duke EPESE substudy entitled "Race Differences in the Prevalence and Incidence of Dementia". This study is significant because it is the first to address the risks associated with the use of specific medication classes and cognitive impairment in a representative sample of community dwelling elders and it also compares methods to assess medication appropriateness at both the population and the patient level. The ultimate objective of this proposal is to enhance the health of the cognitively impaired elderly by improving medication appropriateness.